The rejection of polar organic solutes in aqueous solution by an interpolymer anionic composite reverse osmosis membrane

Reverse osmosis separation for many kinds of polar organic solutes (alcohols, phenols, monocarboxylic acids, amines, and ketones) was examined by an anionic charged composite membrane. The solute permeation was carried out in single-solute aqueous solution (200 mg/L) under applied pressure of 7.88 MPa at 25°C. The correlation between the solute rejection and polar parameters for these organic solutes have been investigated. For n-alkyl alcohols, monocarboxylic acids, and ketones, the solute rejection increases with molecular weight and/or molecular branching. For undissociable polar organic solutes such as alcohols and ketones, solute rejections are closely related with the Taft's number. For dissociable polar organic solutes, solute rejections depend greatly upon the dissociation constant and the degree of dissociation of solute. This membrane showed higher rejection (80%) for phenol in an undissociated state at a 98% rejection level of NaCl. Also, rejections of phenolic derivatives depend upon the pH value of the feed solution and the polar effect of substituted groups. For acetic acid and methylamine, the solute rejection increases proportionally to the degree of dissociation of solute. From these facts, the main factors in reverse-osmosis separation by an anionic composite membrane are discussed.     

Removal of API's (Active Pharmaceutical Ingredients) from Organic Solvents by Nanofiltration

Abstract

The removal of 5 specific active pharmaceutical ingredients (API's) with molecular weight of 189, 313, 435, 531, and 721, respectively, from toluene, methylene chloride, and methanol was studied by using solvent resistant nanofiltration. Three membranes of the StarMem series (120, 122, and 228), with cut‐off values of 200, 220, and 280 respectively, were used in the experiments. Although the rejections expected from the size difference between solutes and membrane pores are high, the results largely depended on the solvent used. For toluene, rejections were rather small, due to the low molecular weight of the solutes of interest (all API's except for the largest compound). Modelling of the rejection curve showed that the minimum molecular weight of a solute to obtain a rejection of 90% in toluene with the membranes used, is ca. 600. The application in methylene chloride was unsuccessful due to partial dissolution of the membrane top layer; other polymeric membranes such as the Solsep series might be more successful. The rejections in methanol were sufficiently high (>90%) to allow implementation: the rejection can be significantly increased by using a module design with double membrane passage and recirculation of the retentate, as was calculated from mass balances. A comparison of a (single pass) nanofiltration system with a throughput distillation unit, currently in use, showed that the energy consumption is 200 times lower in the nanofiltration system.     

Solute transport in non‐aqueous nanofiltration: effect of membrane material

Nanofiltration experiments in methanol and ethanol were carried out for six reference components with different molecular weights (MW 228–880) and polarities (logP 0–12). The contribution of diffusion to solute transport, calculations based on results from cell diffusion experiments, was found to be only 1–7%; solute transport occurs mainly by convection. Furthermore, it was found that solute transport is influenced by solute–solvent–membrane interactions. Solvent–solute interactions (solvation) cause a different effective solute diameter in each solvent: it is smaller in ethanol than in methanol, resulting in lower rejections in ethanol than in methanol. Solute rejection increases with increasing molecular size (for components with similar polarity). Solute–membrane interactions were expressed in polarity terms and charge effects. A decrease of the rejection with decreasing solute polarity (for components with similar MW) was observed. Since non‐polar components (high logP) are exposed to smaller repulsion forces from the polymeric membrane material, the resistance against solute permeation is lower for these components. The solvent–membrane interactions were found to result in solvation of the pore wall; the degree of membrane solvation is different for each solvent. It is determined by the affinity between the solvent and the membrane, and by the molecular size of the solvent. In ethanol, hydrophilic membranes show a larger drop in solute rejection than hydrophobic membranes. The differences in solvent–membrane affinity (measured by contact angle) are much smaller for the first membranes, and therefore pore wall solvation decreases with increasing solvent size. Hydrophobic membranes have a much larger affinity for ethanol than for methanol, leading to stronger interactions, but undergo competitive forces due to the larger solvent size. Therefore, the difference in degree of solvation and effective pore diameter is less pronounced. Based on these three observed or postulated interactions, rejections of all six reference solutes in methanol and ethanol could be explained. Copyright © 2005 Society of Chemical Industry     

Reverse osmosis of lactic acid fermentation broths

Lactic acid model solutions and fermentation broths were concentrated using a tubular thin-film composite reverse osmosis membrane. Flux increased linearly with applied transmembrane pressure and was relatively unaffected by flow rate. Osmotic pressures of 1% lactate solutions were 280–560 kPa, depending on the pH or degree of dissociation. Rejections increased with applied pressure. Higher pH caused a slight decrease in flux (due in part to the higher osmotic pressure) and a significant increase in rejection. Above pH 5·6, rejections of lactate and residual sugars were > 97%. In contrast, with cellulose acetate membranes, flux was generally lower and lactate rejection was proportional to the degree of dissociation at lower pressures.     

A comparative study for boron removal from seawater by two types of polyamide thin film composite SWRO membranes

In this paper, field performance of a small-scale seawater reverse osmosis unit installed in Urla Bay-Izmir, Turkey was analyzed and presented. The design of SWRO system in Urla consists of two types of FilmTec polyamide thin film composite spiral wound seawater reverse osmosis membranes (high rejection FILMTEC XUS SW30XHR-2540 RO membrane and FILMTEC SW30-2540 RO membrane) which could be operated in parallel. To make a comparative study between two types of membranes regarding their desalination performances and boron rejections, each membrane was operated individually for each set of experiments. This comparison was made via investigation of the effects of feed seawater temperature (10–16 °C), operating pressure (55, 60 and 62 bar), and pH adjustment on the feed side (pH 7.0–7.5).     

Reverse osmosis separations for some alcohols and phenols in aqueous solutions using aromatic polyamide membranes

The performance of aromatic polyamide membranes for reverse osmosis separations of eight alcohol and four phenol solutes in dilute aqueous solutions has been studied. The Taft polar parameter σ^* for the solutes studied were in the range of ?0.3 to 1.388. Positive solute separations were obtained for each one of the solutes. In the σ^* value range of ?0.3 to 0, data on PR/PWP ratio scattered close to 1, and solute separation decreased with increase in σ^*. For the phenol solutes, PR/PWP ratio decreased and solute separation increased with increase in σ^*. The results are interpreted on the following basis. The aromatic polyamides are more nonpolar than cellulose acetates. In the σ^* range of ?0.115 to ?0.3, solute separation is governed primarily by polar interactions; in this range, solute transport parameter D_AM/Kδ is well correlated by the expression D_AM/Kδ = C^* exp (ρ^*σ^*). The solute separation for ethyl and methyl alcohol solutes (σ^* = ?0.1 and 0, respectively) is reduced by the nonpolar character of the membrane material. Positive solute separation for each of the phenolic solutes is due to preferential sorption of solute at the membrane-solution interface caused by both the nonpolar character of the membrane material and acidity of the solutes.     

反渗透、纳滤过程的物理化学研究(Ⅰ)多孔膜的溶质脱除率方程和膜渗透通量方程

引言 溶解扩散模型、摩擦模型、Sourirajan方程和表面力孔流模型均假设稳态条件下溶质通量和溶液体积通量恒定[13].     

RO membrane solute rejection behavior at the initial stage ofcolloidal fouling

This study investigated the rejection of salt and inert organic compounds by reverse osmosis membranes during the initial stage of colloidal fouling. Results of laboratory-scale experiments showed that colloidal fouling caused a marked decrease in flux, salt rejection and rejection of organics with molecular weight (MW) smaller than about 100 g/mol. Removal of neutrally charged organics was mainly by size or steric exclusion. Rejection of xylose, which has MW >100 g/mol, was not affected much by colloidal fouling. The decrease in salt and low MW organic rejections during the initial stage of colloidal fouling was attributed to cake-enhanced concentration polarization, whereby the colloidal cake layer hindered back diffusion of solutes from the membrane surface to the bulk solution, resulting in higher solute concentration gradient across the membrane. At higher channel wall shear rate, the rates of colloidal deposition, flux decline, decrease in salt rejection, and decrease in low MW organic rejection were lower.     

Assessment of a semi‐quantitative method for estimation of the rejection of organic compounds in aqueous solution in nanofiltration

A large number of different mechanisms describing the retention of dissolved organic compounds in nanofiltration have been proposed. A recent review identified the parameters possibly involved in the separation performance and suggested a qualitative classification of dissolved compounds. Continuing this approach, a semi‐quantitative assessment of the observed rejections in nanofiltration is given in this paper, based on threshold values of key parameters such as molecular weight and molecular weight cut‐off (MWCO), molecular size, pH and pK_a, hydrophobicity (logK_ow) and membrane charge. Experimental values and literature data were used to provide a broad basis for comparison. It was concluded that (a) all categories that contain hydrophobic components are badly defined, in particular for small components, with rejections varying from low to very high, (b) all components that contain hydrophilic components have relatively high rejections and (c) all categories that contain charged components have well‐defined, high rejections (intermediate for membranes with low surface charge). In all cases, the average rejection is higher when the component's molecular weight is larger than the MWCO of the membrane and when the molecular size is larger than the pore size of the membrane. Copyright © 2006 Society of Chemical Industry     

A pilot scale study of reverse osmosis for the purification of condensate arising from distillery stillage concentration plant

Reverse osmosis (RO) is an interesting process to eliminate small organic solutes (carboxylic acids and alcohols) from distillery condensates before recycling them into the fermentation step. This work investigates the influence of transmembrane pressure, pH and volume reduction factor (VRF) on the efficiency of reverse osmosis treatment of condensate from distillery stillage concentration at pilot scale using three pre-selected membranes (CPA2 and ESPA2 from Hydranautics, BW30 from DOW). Performances were assessed according to permeate flux, solutes rejection and abatement of fermentation inhibition. Transmembrane pressure increase leads to an increase of these three parameters with a plateau for rejections and abatement at 20 bar; however, in order to comply with membranes manufacturer's recommendations and to limit or delay polarization and fouling, it was decided to keep the permeate flux below a value of 30 L h⁻¹ m⁻². This corresponded to a maximum pressure of 10 bar for CPA2 and ESPA2 membranes and 25 bar for BW30 membrane. pH increase leads to a diminution of permeate flux and an increase of carboxylic acids rejection whatever the membrane; nevertheless, no abatement of fermentation inhibition is observed. Increasing VRF provokes a decrease of the permeate flux. Although local rejections are stable, the mean rejection assessed with the raw condensate (feed) and the mean permeate decreases. However, the fermentation inhibition remains under 10% up to a VRF of 8. BW30 membrane exhibits the highest rejections and inhibition abatement. On the basis of the pilot scale results with the BW30 membrane, a preliminary estimation of the membrane area is proposed for an industrial plant with 100 m³ h⁻¹ of condensate flow rate and the optimized parameters (pressure 25 bar, no pH modification, VRF 4 and 8).     

Performance of Nanofiltration Membranes in Ethanol

Abstract

Several nanofiltration membranes were tested for flux and rejection of selected solutes in ethanol. The membranes were initially conditioned with pure solvent containing increasing concentrations of ethanol. Flux decreased with increase in ethanol concentration and increased at higher temperatures and pressures. The type of solute had an influence on membrane rejection profiles. The DK membrane showed increasing rejection of polyethylene glycols (PEG) dissolved in ethanol from 29% at a molecular weight (MW) of 200 to 80% at MW 1000. However, the MW of sugars and lipids had little or no effect on rejection with the DK membrane; their rejection averaged 87%. In contrast, the TFC‐SR1 membrane showed higher rejections with higher MW compounds: lipid rejection increased from 19% to 71%, sugars from 35% to 85%, and lipids from 77% to 89%. The TFC‐SR2 membrane was much more open and showed the lowest rejections of all these compounds. Flux generally showed opposite trends, with the DK showing the lowest flux and the SR2 the highest.     

Membrane Concentration and Separation of L-Aspartic Acid and L-Phenylalanine Derivatives in Organic Solvents

Abstract

The concentration and separation of the amino acids N-benzyloxycarbonyl L-aspartic acid and L-phenylalanine methyl ester hydrochloride in organic solvents have been investigated using reverse osmosis membranes of two types of cellulose acetate, a nanofiltration membrane of polyamide-polyphenylene sulfone (PA-PPSO) composite and a gas separation membrane of polyimide composite in a stirred batch cell. The organic solvents used included primary, secondary, and tertiary alcohols, an ester, and a ketone. There were significant variations in permeate flux, solute rejection, and membrane stability. Usually the rejection of both amino acids was similar; however, certain membrane-solvent combinations gave significantly different levels of rejection. The highest rejection of amino acids (～0.94) at the lowest pressure of 0.5 MPa was obtained with the PA-PPSO membrane using methanol as a solvent. The cellulose acetate membranes gave reasonable rejection and fluxes but the membrane stability was very poor. The performance of the polyimide composite membrane was good with ethanol but poor with other solvents. The PA-PPSO membrane with methanol as solvent appeared the most promising combination, and the separation performance according to concentration polarization was discussed.

     

Reverse osmosis rejection by hydrous inorganic precipitate–cellulose composite membrane

Composit membranes were prepared by impregnating hydrous inorganic precipites (Fe and Cr oxides, nickel chromate or molybdate) into a cellulose acetate membrane which serves as a support. In these membranes, the hydrous iron oxide composite membrane showed high reverse osmosis properties. The permeability of the composite membrane did not decline with time, and the rejection of organic solute was not appreciably affected by impregnation. However, the rejection of electrolyte increased with impregnation. The measurements of membrane potential revealed that salt rejection is primarily attributable to the effect of membrane charge, i.e., to ion exclusion effect. In electrolyte solutions of 1:1 and 2:1, the membrane was anion selective; on the other hand, in electrolyte solution of 1:2, the membrane was cation selective. These results are characteristic of impregnated hydrous metal oxide.     

基于吸附-扩散模型和遗传算法的反渗透膜性能预测

基于反渗透膜的吸附-扩散模型,导出了多孔膜对溶质的浓缩比表达式和组件的溶质平均脱除率。采用遗传算法对膜的性能进行了数学模拟,计算出了氯化钠对ESPA2反渗透膜的膜相溶质分配系数、溶质的膜相扩散系数、溶剂(水)的膜相扩散系数、膜的孔隙率和膜分离层厚度,计算得到的膜分离层厚度和膜相中水扩散系数等参数与芳香聚酰胺复合膜的测定值相吻合,其中,氯化钠透过率的均方差在10-7～10-6之间。     

Reverse osmosis separation of hydrocarbons in aqueous solutions using porous cellulose acetate membranes

A physicochemical parameter, represented by the symbol Σs^*, based on molar solubility in water and molar attraction constants of Small, has been developed to express quantitatively the relative hydrophobicity, or nonpolar character, of the hydrocarbon molecule. The value of Σs^* can be calculated for a hydrocarbon from its chemical structure. The scale of Σs^* is consistent within each group of aromatic, cyclic, and noncyclic hydrocarbons. Reverse osmosis data have been obtained at 250 psig for single-solute aqueous feed solution systems involving low concentrations of 39 different hydrocarbons (including 13 aromatics, 10 cyclic, and 16 noncyclic compounds) and several samples of cellulose acetate membranes of different surface porosities. The effect of operating pressure on membrane performance has also been studied for two aromatic hydrocarbon solutes. The values of Σs^* for the solutes used were in the range of 425 to 924 for aromatic hydrocarbons, 521 to 931 for cyclic hydrocarbons, and 369 to 960 for noncyclic hydrocarbons. The reverse osmosis data have been correlated with Σs^* for each group of hydrocarbons studied. In all cases, positive solute separations were obtained, and the ratio [PR]/[PWP] was less than 1. With respect to each film, solute separation increased with increase in Σs^*, and decreased with increase in operating pressure. Also, solute separation decreased in the order aromatic hydrocarbon > cyclic hydrocarbon > noncyclic hydrocarbon at any given value of Σs^*. At a given operating pressure, for low values of Σs^* (～500 or less) solute separation increased with progressive decrease in average pore size on the membrane surface. For high values of Σs^* (～800 or more), solute separation initially increased with decrease in average pore size, then passed through a maximum and minimum with further decrease in average pore size, and again increased with still further decrease in average pore size. The results are discussed on the basis of preferential sorption of solute at the membrane–solution interface under the experimental conditions studied.     

Effect of Cyclic Changes in Temperature and Pressure on Permeation Properties of Composite Polyamide Seawater Reverse Osmosis Membranes

The effects of cyclic changes in feed water temperature and pressure on permeate flux, solute rejection, and compaction in spiral wound composite polyamide seawater reverse osmosis membranes were examined with pure water and 4% NaCl solutions. A membrane permeability hysteresis or memory effect due to the up and down temperature and pressure sequences was only seen with the saline water studies. However, the observed changes appeared to be reversible and were consistent with the Spiegler-Kedem/ Film Theory and the Kimura-Sourirajan Analysis/ Film Theory models. The overall results suggest that the net effect on permeance and solute rejection is the consequence of several interactions with feed/operating temperatures affecting membrane porosity and water/solute cluster size, and transmembrane pressure influencing membrane compaction.     

Reverse osmosis separation of phenols in aqueous solutions using porous cellulose acetate membranes

Reverse osmosis separations of phenol (9.4 to 108 ppm), p-cresol (108 ppm), and p-chlorophenol (129 ppm) were studied using Loeb-Sourirajan-type porous cellulose acetate membranes, and single-solute aqueous feed solutions at 500 psig and the indicated solute concentrations. It was found that, by dissociating the solute by changing the pH of the feed solution, all the above phenols could be separated by reverse osmosis. Solute separation increased with increase in the degree of dissociation of the solute in the feed solution; and, by the appropriate choice of pore size on the membrane surface, separations of phenol approaching the degree of dissociation of phenol in the feed solution could be obtained under the operating conditions used. Similar experiments using aniline (93 ppm) as the solute showed that dissociation of solute molecules in the feed solution could be a technique generally applicable for the reverse osmosis separation of nonionic solutes in aqueous solution. The effects of operating pressure in the range 250 to 1500 psig and pore size on the membrane surface on the separation of un-ionized phenol and p-chlorophenol showed that, with respect to single-solute aqueous feed solutions of phenols, the component whose relative acidity was greater was preferentially sorbed at the cellulose acetate membrane—aqueous solution interface, and the solute concentration in the membrane-permeated product solution was a function of the extent and mobility of each of the sorbed species.     

Rejection of As(III) and As(V) from arsenic contaminated water via electro-cross-flow negatively charged nanofiltration membrane system

A specially designed electro-cross-flow nanofiltration (NF) membrane system was used for this investigation. To enhance the rejection of arsenic ionic species like H₂AsO₄⁻, a NF membrane having a negative surface charge was fabricated via the interfacial polymerization process. The membrane was characterized by SEM, AFM, surface charge density, molecular weight cut-off (MWCO), total and skin thickness and pure water flux. The parameters that affected the rejections of As(III) and As(V) were studied; they included the initial arsenic concentration, the applied potential, pH of the feed, the cross-flow filtration pressure and the presence of different salts in the feed. Among those parameters, the pH of the feed greatly affected As(V) rejection; As(V) ([As(V)]_o = 1000 ppb) rejection was increased from 72.3 to 98.5% when pH of the feed was changed from 3.0 to 10.0. This might be due to the fact that higher pH enhanced the formation of negative divalent anion like HAsO₄²⁻ which should be rejected more effectively by the negative surface charge of the NF membrane. Beside the effect of the negative surface charge of the membrane, applied potential increased the As(V) rejection by 48.2% when the applied potential was increased from 0 to 2.0 V for a feed containing 1000 ppb initially. For the same change of applied potential rejection of As(III) was increased from 52.3 to 70.4%; this might be the result of the formation of anionic species like H₂AsO₃⁻ from the neutral molecule of H₃AsO₃ by the applied potential.     

The effect of chelation on the selective transport of alkaline earth metal ions in asymmetric cellulose acetate hyperfiltration membranes. II

The effects of chelation on the transport of calcium and magnesium, both separately and in a variety of admixtures, in a controlled series of asymmetric cellulose acetate membranes were characterized. Ethylenediaminetetraacetic acid (EDTA) and ethylenebis(oxyethylenenitrilo)tetraacetic acid (EGTA) were used as chelating agents for the alkaline earth metal ions. Asymmetric cellulose acetate membranes annealed at 70°, 75°, and 85°C were studied. Chelation of each of these alkaline earth metals ions in aqueous solutions at pH 6, by either EDTA or EGTA, significantly increased the overall hyperfiltration rejections of these metals by all the membranes studied. The increase in rejection varied montonically with the fraction of metal ion complexed. The higher rejection of metal chelates, compared to the rejection of unbound metal ions, was considered to be the result of the significantly larger size of the chelated species. Calculations suggested that selective (or competitive) chelation took place at pH 6 in a mixture of calcium and magnesium ions in the presence of a stoichiometrically limiting amount of chelating agent. Calcium successfully competed for most of the available chelating agent in equimolar aqueous solutions of chelating agent, calcium, and magnesium. The calcium rejection was explained primarily in terms of the effects of chelation per se on the effective size of the formed complex even in feeds comprised of these ternary solute mixtures. The complexation reaction between magnesium and EGTA is, however, so unfavorable at pH 6 that the Mg²⁺ ion remains uncomplexed even in the presence of an equivalent amount of EGTA. The observed increased rejection of magnesium ions, therefore, in ternary systems was explained by electroneutrality criteria and by solute–membrane interactions involving the various calcium species and the membranes.     

Reverse osmosis separation of single and mixed alcohols in aqueous solutions using porous cellulose acetate membranes

Reverse osmosis transport for alcohol-water systems in the Taft number (σ^*) region of 0 to ?0.3 is explored in detail. The numerical value of the polar functional constant for alcohols is 15.5 for the above σ^* region in the operating pressure range of 50 to 500 psig for the cellulose acetate membrane material used. An analysis of the combined effect of operating pressure and mass transfer coefficient on the high-pressure side of the membrane shows that, under certain conditions, solute separation could pass through a maximum with increase in operating pressure. A general experession for solute separation is derived as a function of pore structure on membrane surface, polarity of solute, and operating conditions of the experiment. Alcohols behave independently in mixed solute systems. A method is described and illustrated for predicting alcohol separation in alcohol–sucrose–water feed solutions from data on single solute systems.